Brian P. Flaherty

An Alternative Framework for Defining Mediation.

The present article provides an alternative framework for evaluating mediated relationships. From this perspective. a mediated process is a chain reaction, beginning with an independent variable that affects a mediator that in turn affects an outcome. The definition of mediation offered here, presented for stage sequences, states three conditions for establishing mediation: (a) the independent variable affects the probability of the sequence no mediator to mediator to outcome; (b) the independent variable affects the probability of a transition into the mediator stage; (c) the mediator affects the probability of a transition into the outcome stage at every level of the independent variable. This definition of mediation is compared and contrasted with the well-known definition of mediation for continuous variables discussed in Baron and Kenny (1986), Judd and Kenny (1981), and Kenny, Kashy, and Bolger (1997). The definition presented in this article emphasizes the intraindividual, time-ordered nature of mediation.

Metabolic Control and Quality-of-Life Self-Assessment in Adolescents With IDDM

OBJECTIVE To examine the relation between metabolic control and self-assessed quality of life in adolescents with IDDM. RESEARCH DESIGN AND METHODS The Diabetes Quality of Life (DQOL) questionnaire for youths was given to 69 subjects with IDDM aged 10-20 years at the time of their outpatient visit. Subjects with IDDM of < 1 years duration or with documented psychotic disorder or mental retardation were excluded. Metabolic control was assessed by the mean HbA1c during the preceding year (long-term), by a single HbA1c at the time of the visit (short-term), and by the number of acute events related to IDDM in the preceding year. RESULTS The DQOL score correlated with mean HbA1c (β = 6.13, R2 = 0.22, P = 0.0122) and single HbA1c (β = 3.94, R2 = 0.18, P = 0.05). Self-health assessment was the best predictor of DQOL score (β = −44.42, R2 = 0.45, P < 0.0001). The Worries subscale score on DQOL correlated with the occurrence of acute events (β = 6.97, R2 = 0.2, P = 0.006), but did not correlate with either HbA1c level. Correlations of mean HbA1c with the predictors were stronger than the correlations of single HbA1c with the same predictors. CONCLUSIONS Metabolic control and quality of life are two important outcomes of IDDM care. In our study, adolescents in better metabolic control report better quality of life. Both components need to be addressed in developing successful diabetes treatment strategies for adolescents with IDDM.

Dynamic Patterns of Time Use in Adolescence

Patterns of time use are tangible representations of individual identity and the meaning of age groups in the life course. How do young people allocate their time to multiple domains of involvement, including the school, workplace, family, and peer group? Drawing on longitudinal data from the Youth Development Study (N = 1,010), a person-centered analytic strategy was used to describe configurations of time use through the high school years. Over half of the students were engaged in many domains, although a substantial percentage of students focused their time on one or two domains outside the school. Students who were highly engaged in multiple domains tended to remain so across grade levels, whereas students focused on one or two domains frequently changed their commitments. Plans for school, grade point average, future orientations that emphasize marriage and good citizenship, and gender significantly predicted time-use patterns. These findings elucidate connections among school, work, and other contexts through the high school years.

Gendered dimensions of smoking among college students

Ethnographic research, including interviews, focus groups, and observations were conducted to explore gendered dimensions of smoking among low level smokers, including the acceptability of smoking in different contexts; reasons for smoking; the monitoring of self and friends’ smoking; and shared smoking as a means of communicating concern and empathy. Important gendered dimensions of smoking were documented. Although males who smoked were described as looking manly, relaxed, and in control, among females, smoking was considered a behavior that made one look slutty and out of control. Young women were found to monitor their own and their friends’ smoking carefully and tended to smoke in groups to mitigate negative perceptions of smoking. Gender-specific tobacco cessation programs are warranted on college campuses.

Using data augmentation to obtain standard errors and conduct hypothesis tests in latent class and latent transition analysis

Latent class analysis (LCA) provides a means of identifying a mixture of subgroups in a population measured by multiple categorical indicators. Latent transition analysis (LTA) is a type of LCA that facilitates addressing research questions concerning stage-sequential change over time in longitudinal data. Both approaches have been used with increasing frequency in the social sciences. The objective of this article is to illustrate data augmentation (DA), a Markov chain Monte Carlo procedure that can be used to obtain parameter estimates and standard errors for LCA and LTA models. By use of DA it is possible to construct hypothesis tests concerning not only standard model parameters but also combinations of parameters, affording tremendous flexibility. DA is demonstrated with an example involving tests of ethnic differences, gender differences, and an Ethnicity x Gender interaction in the development of adolescent problem behavior.

Regular smokeless tobacco use is not a reliable predictor of smoking onset when psychosocial predictors are included in the model

Tomar analyzed the CDCs Teenage Attitudes and Practices Survey (TAPS) and reported smokeless tobacco may act as a starter product for or gateway to cigarettes. Regular smokeless tobacco users at baseline were said to be 3.45 times more likely than never users of smokeless tobacco to become cigarette smokers after 4 years (95% CI=1.84-6.47). However, this analysis did not take into account well-known psychosocial predictors of smoking initiation. We reanalyzed TAPS to assess whether including psychosocial predictors of smoking affected the smokeless tobacco gateway effect. Experimenting with smoking, OR=2.09 (95% CI=1.51-2.90); below average school performance, OR=9.32 (95% CI=4.18-20.77); household members smoking, OR=1.49 (95% CI=1.13-1.95); frequent depressive symptoms, OR=2.19 (95% CI=1.25-3.84); fighting, OR=1.48 (95% CI=1.08-2.03); and motorcycle riding, OR=1.42 (95% CI=1.06-1.91) diminished the effect of both regular, OR=1.68 (95% CI=.83-3.41), and never regular smokeless tobacco use, OR=1.41 (95% CI=.96-2.05), to be statistically unreliable. Analyzing results from a sample of true never smokers (never a single puff) showed a similar pattern of results. Our results indicate that complex multivariate models are needed to evaluate recruitment to smoking and single factors that are important in that process. Tomars analysis should not be used as reliable evidence that smokeless tobacco may be a starter product for cigarettes.

Assessing reliability of categorical substance use measures with latent class analysis

This article illustrates the use of the latent class model to identify classes of individuals and to assess the psychometric reliability of categorical items. The latent class model is a categorical latent variable model used to identify homogeneous classes of respondents such that class membership accounts for item responses. The assessment of measurement reliability comes directly from the estimates of the model. Although not based on classical test theory, the reliability assessment procedures described here answer the same question-that is, how consistent or dependable is measurement? The goal is to identify reliable indicators of a characteristic by examining measurement error and the inter-relatedness of the items. Methods for estimating the reliability of individual items as well as sets of items are presented. These methods are illustrated with data on cigarette smoking from a national sample of adolescents. By using the procedures described here, researchers are able to determine: (1). which classes of people are measured well and which are not; (2). which items perform well and which do not; and (3). whether items need to be altered or added in order to measure and identify particular classes better.

Chilean adolescents’ beliefs about the legitimacy of parental authority: Individual and age-related differences

Individual and age-related differences in the patterning of adolescents’ beliefs about the legitimacy of parental authority were examined in a sample of 3425 Chilean adolescents (Mage = 15.0). During early, middle, and late adolescence, three analogous patterns of beliefs about the legitimacy of parental authority were identified using latent class analysis (LCA). Youth in the Parental Control class ceded parents legitimate control over issues in the multi-faceted and prudential domains and were relatively likely to cede parental control over the personal domain. Those in the Shared Control class differentiated the prudential from other domains. Those in the Personal Control class denied parents legitimate authority over issues in all domains. Within analogous classes, younger adolescents were more likely to grant parents legitimate authority than older adolescents. Results are consistent with prior research documenting age-related differences, but raise important questions about the normative nature of age-related change in legitimacy beliefs. The advantages of studying sub-groups and variability in the patterning of legitimacy beliefs are discussed.

Understanding Ageing: Further Commentary on the Limitations of Cross-Sectional Designs for Ageing Research

Critical issues regarding the utility of age-heterogeneous cross-sectional samples for estimating associations among rates of change are briefly reviewed and further examined in response to the four insightful commentaries on our original discussion paper. These general issues are considered in terms of current practice in gerontological studies. Several of the commentaries refer to additional problems (e.g., selection/attrition, cohort differences) with cross-sectional designs for inferring correlated rates of change, often modeled as shared age-related variance. The commentators concurred with, and performed simulations that support, our primary contention that the influence of mean population trends over time has a major confounding influence on cross-sectional estimates of association. While this confound has been acknowledged and demonstrated in a number of publications over this century, cross-sectional studies and covariance modeling remain common practice in gerontological studies. We question the value of this practice for advancing aging theory and would like to see greater emphasis placed on longitudinal studies and the use of cross-sectional designs and analytic methods that permit separation of age-related mean trends from estimates of association between age-related variables.

Do wealth disparities contribute to health disparities within racial/ethnic groups?

Background Though wide disparities in wealth have been documented across racial/ethnic groups, it is largely unknown whether differences in wealth are associated with health disparities within racial/ethnic groups. Methods Data from the Survey of Consumer Finances (2004, ages 25–64) and the Health and Retirement Survey (2004, ages 50+), containing a wide range of assets and debts variables, were used to calculate net worth (a standard measure of wealth). Among non-Hispanic black, Hispanic and non-Hispanic white populations, we tested whether wealth was associated with self-reported poor/fair health status after accounting for income and education. Results Except among the younger Hispanic population, net worth was significantly associated with poor/fair health status within each racial/ethnic group in both data sets. Adding net worth attenuated the association between education and poor/fair health (in all racial/ethnic groups) and between income and poor/fair health (except among older Hispanics). Conclusions The results add to the literature indicating the importance of including measures of wealth in health research for what they may reveal about disparities not only between but also within different racial/ethnic groups.